PREDICTABILITY BETWEEN OBJECTIVE PHYSICAL FACTORS OF FABRICS AND SUBJECTIVE PREFERENCE VOTES FOR DERIVED GARMENTS

The relationships have been investigated between subjective preference votes, collected under three different conditions, for garments derived from a range of fabric materials and objective physical factors (where "factors" are grouped properties) measured for the same materials. The technique of canonical correlation analysis was applied, and three canonical correlations between subjective preference votes and objective physical factors were found to be highly significant (P <0.0001, r > 0.950). Canonical redundancy analysis showed that the objective physical factors of fabrics had great predictive power for the subjective votes, with cumulative redundancy over 0.983. The subjective preference votes, however, were poor predictors for the physical factors of fabrics, with cumulative redundancy only 0.467. The squared multiple correlations suggested that the first three canonical variables of the objective factors of fabrics had very good predictive power for all three subjective preference votes, but the first three canonical variables of the subjective preference votes showed good predictive power only for fabric roughness and fullness and for fabric wettability, not for other objective factors.